Update app.py

app.py

@@ -85,3 +85,56 @@ vectordb = Chroma.from_documents(documents=all_splits, embedding=embeddings, per
 
 # specify the retriever
 retriever = vectordb.as_retriever()
+
+# build huggingface pipeline for using zephyr-7b-alpha
+pipeline = pipeline(
+        "text-generation",
+        model=model,
+        tokenizer=tokenizer,
+        use_cache=True,
+        device_map="auto",
+        max_length=2048,
+        do_sample=True,
+        top_k=5,
+        num_return_sequences=1,
+        eos_token_id=tokenizer.eos_token_id,
+        pad_token_id=tokenizer.eos_token_id,
+)
+
+# specify the llm
+llm = HuggingFacePipeline(pipeline=pipeline)
+
+# build conversational retrieval chain with memory (rag) using langchain
+def create_conversation(query: str, chat_history: list) -> tuple:
+    try:
+
+        memory = ConversationBufferMemory(
+            memory_key='chat_history',
+            return_messages=False
+        )
+        qa_chain = ConversationalRetrievalChain.from_llm(
+            llm=llm,
+            retriever=retriever,
+            memory=memory,
+            get_chat_history=lambda h: h,
+        )
+
+        result = qa_chain({'question': query, 'chat_history': chat_history})
+        chat_history.append((query, result['answer']))
+        return '', chat_history
+
+
+    except Exception as e:
+        chat_history.append((query, e))
+        return '', chat_history
+
+# build gradio ui
+with gr.Blocks() as demo:
+    
+    chatbot = gr.Chatbot(label='Chat with your data (Zephyr 7B Alpha)')
+    msg = gr.Textbox()
+    clear = gr.ClearButton([msg, chatbot])
+
+    msg.submit(create_conversation, [msg, chatbot], [msg, chatbot])
+
+demo.launch()